Abstract: For Out Of Sequence Measurement（OOSM） problem in centralized multi-sensor target tracking system, an algorithm based on optimal fusion has been proposed, which is based on the idea of weighted covariance fusion and in the sense of the trace of fusion estimation error covariance matrix smallest. The algorithm aligns the OOSM to the time of the latest state estimate, and then it is fused with the latest state estimate using the covariance weighted method. In order to calculate the noise correlation of OOSM and the other measurements, equivalent measurement is introduced. Theoretical analysis and simulation show that the algorithm can effectively deal with the problem of OOSM multi-step delay. Its performance is almost optimal and decreases small as delay steps increases, and it has the same filtering accuracy with data saving method with the optimal performance and a small amount of additional storage.

Key words: data fusion, out-of-sequence-measurement, covariance weighting, multiple-step delay, equivalent measurement

摘要： 针对集中式多传感器目标跟踪系统中存在的无序量测问题，基于协方差加权融合的思想，在融合估计误差协方差矩阵迹最小意义下，建立了基于最优融合的多步延迟无序量测更新算法。该算法先将无序量测配准到最新状态估计的时刻，将其与之进行协方差加权融合。为进行无序量测与各传感器量测噪声相关性的计算，引入了等效量测。通过理论分析和仿真实验说明该算法能有效处理无序量测多延迟问题，其性能接近最优且随延迟步数增加性能下降非常小，而且有与最优的数据缓存法相同的滤波精度，以及较小的额外存储量。

关键词: 数据融合, 无序量测, 协方差加权, 多步延迟, 等效量测